2026-10-12T19:03:45Z https://api.figshare.com/v2/oai

oai:figshare.com:article/34033380 2026-09-30T15:13:05Z category_4 category_873 category_734 category_931 category_132 portal_63 item_type_6 month_year_09_2026

GMSF: A Dual-Path Multimodal Framework for Enzyme Function Prediction via Difference Graph Encoding and Multiscale Semantic Fusion Xin Zhao (71840) Haoshu Chen (25144620) Tao Zhang (43681) Yahui Cao (13285995) Haotong Li (13893772) Zhuoran Song (25144623) Bingzhi Li (3398189) Shuo Zheng (458517) Biochemistry Chemical Sciences not elsewhere classified Biological Sciences not elsewhere classified Information Systems not elsewhere classified Infectious Diseases world metabolic networks retrosynthetic pathway planning path multimodal framework number prediction remains multiscale sequence encoding exact enzyme selection enzyme function prediction ecreact dataset demonstrate difference graph encoding coenzyme dependencies prior chemical bond evolution biocyc database validates aligned node mappings level state transitions treat reactions successful application static snapshots reaction centers modality representations immense potential green chemistry explicit atom experimental results experimental implementation critical bottleneck central role catalytic mechanisms 36 %. Enzymes play a central role in green chemistry and biomanufacturing. However, precise enzyme selection via Enzyme Commission (EC) number prediction remains a critical bottleneck between the retrosynthetic pathway planning and experimental implementation. Existing computational methods predominantly rely on single-modality representations that either lose spatial topological structures or treat reactions as static snapshots, failing to capture the explicit atom-level state transitions in chemical bond evolution. To address these limitations, we propose GMSF, a dual-path multimodal deep learning framework. GMSF features two core innovations: Difference Graph Encoding, which leverages atom–atom mapping (AAM) to compute node feature differences and extract structural difference signatures of reaction centers; and Multiscale Sequence Encoding, which uses AAM-aligned node mappings to hierarchically capture chemical semantics. Experimental results on the ECREACT dataset demonstrate that GMSF achieves an accuracy of 93.17% in the highly challenging level 3 (subsubclass) enzyme function prediction, outperforming the state-of-the-art (SOTA) baseline by a margin of 1.36%. Furthermore, its successful application to real-world metabolic networks in the BioCyc database validates its immense potential for deciphering complex biochemical reactions, serving as a powerful computational screening tool to narrow down catalytic mechanisms and coenzyme dependencies prior to the exact enzyme selection. 2026-09-30T00:00:00Z Text Journal contribution 10.1021/acs.jcim.6c02672.s001 https://figshare.com/articles/journal_contribution/GMSF_A_Dual-Path_Multimodal_Framework_for_Enzyme_Function_Prediction_via_Difference_Graph_Encoding_and_Multiscale_Semantic_Fusion/34033380 CC BY-NC 4.0